178 MB. a. BENTHAM OX THE GENEBA 



now carefully examined all the twenty-three species known, most 

 of them from several specimens and in different states, from the 

 young bud to the fully-opened flower, which makes a considerable 

 difference in the apparent arrangement of the stamen^. In all I 

 find, as in Chamcelauciunty Verticordia, and other allied genera, the 

 ten stamens and the ten alternating staminodia, when narrow, 

 yery shortly dilated at the base, and forming a single ring on the 

 margin of the adnate disk, within the petals, but without any 

 overlapping at the base in the bud. The filaments, as in all 

 Myrtacese, are turned down inwards in the bud, and as the an- 

 thers, before the flower opens, are too broad to pack in a single 

 row, we always find them alternately in two rows, one more m- 

 flexed and consequently lower down and inside the other; and 

 it is always, contrary to what we might theoretically expect, the 

 sepaline stamens (those opposite the calyx-lobes) that are the 

 lowest and innermost, the petaline stamens less inflected and 

 more outside ; and the staminodia, still more erect, although usually 

 in the same ring at the base, appear outside of all. Wherever 

 the staminodia are broad, they are even at the base outside of the 

 sepaline or even of both sets of anthers. It follows that they 

 have always a tendency to be nearer the petaline than the sepaline 

 stamens, although in a degree often almost or quite imper- 

 ceptible, but exceedingly variable in different or sometimes in the 

 same species. This tendency is very perceptible in Darwinia 

 fascicular 18^ Eudge, much more so in Hedaroma pinifolia^ Lindl. 

 (which Schauer refers to Qenetyllis, notwithstanding its close re- 

 semblance to D , fascicularis , siiii although the broad staminodia 

 are actually outside the sepaline stamens from the base), and, to 

 ¥ny eyes ait least, imperceptible in D. taxifolia^ A. Cunn., retained 

 in Darwinia by Schauer probably on account of its geographical 

 station. We feel obliged, therefore, to reduce the whole genus 

 Genetyllis to Darvdniuj — a course to be regretted on account of tbe 

 disturbance of nomenclature it occasions, but Eudge's name hfl» 

 undoubtedly the right of priority ; his character was indeed faultyi 

 but so also was De Candolle*s, so that neither genus was at firs^' 

 recognized by Lindley or Endlicher. 



Ferdinand Mueller has more recently, in the Fourth Volume 

 of his ' Fragmenta,' transferred his group Schuermannia to Cha- 

 m€dlatcetumy adding two new and very remarkable species. 1* 

 appears to me, however, that the anthers afford a much more 

 constant distinctive character between Darwinia and Chamal^^^ 



ace 



former 



